A New Method on Uncalibrated Visual Servo Based on Image for Dynamics Model of Robot Manipulators ⋆

For uncalibrated visual servoing for image-based dynamic control of robot manipulators, acquiring an image Jacobian matrix (interaction matrix) or its inverse matrix is necessary and difficult. In this paper, a new controller of uncalibrated visual sevro using pseudo inverse of image Jacobian matrix is presented, which based on dynamics of robot manipulators. In this method, camera’s intrinsic and extrinsic parameters are not known; joints’ speed of robot is non-essential. On the basis of the nonlinear robot dynamics, asymptotic convergence of the image errors to zero by the Lyapunov theory is proved. Simulations have been conducted to verify performances of the proposed controller. Results demonstrate the good convergence of image errors.